Fly embryo nuclei detect a 10% difference in Bicoid (blue) concentration that either does or doesn't activate the head gene hunchback (green).

GREGOR

Running from head to tail in a fly embryo, the Bicoid (Bcd) gradient is the blueprint for anterior–posterior development. How cells read the blueprint to give a precise pattern was thought to require multiple mechanisms to smooth out sloppy Bcd input signals. Now, two papers from Thomas Gregor and colleagues (Princeton University, Princeton, NJ) put numbers on the cells' Bcd readout abilities and show the system to be highly precise—approaching the limits set by the inherent noise in any physical system.

Using live embryos to image the dynamics of the Bcd gradient, the team determined that the gradient was established within ∼1 hour after fertilization and that Bcd diffused through the cytoplasm of the syncytial embryo with a diffusion constant of 0.3 μm2 per second. But if one assumes that simple diffusion establishes the gradient, Bcd would never reach steady-state within the developmental timeframe. More work is needed to find other mechanisms that are at play.

This first look at a transcription factor's behavior in a live organism also revealed tightly regulated levels of nuclear Bcd between mitotic cycles. During four syncytial cycles, when nuclei multiply rapidly and get smaller, the Bcd concentration in a given nucleus returned at each interphase to within 10% of its starting concentration, holding the blueprint coordinates steady.

At the midpoint of the embryo—where Bcd levels are at the head–tail borderline—nuclei held ∼700 molecules of Bcd. A precision of 10% thus means that midpoint cells detected a difference of ∼70 molecules. The noise in Bcd readout (measured by its activation of the head gene hunchback) was also 10%, as was the reproducibility of the Bcd gradient from embryo to embryo.

The work argues that the cells along the embryo's anterior–posterior axis determine their position by a precise readout of their own Bcd concentration to either activate hunchback or not. And, the authors note, the readout may be even more exact, since the repeated 10% figure is “disturbingly close” to the noise introduced by their instrumentation.

References:

Gregor, T., et al.
2007
.
Cell.
130
:
141
–152.

Gregor, T., et al.
2007
.
Cell.
130
:
153
–164.